AU2021203706A1

AU2021203706A1 - Updating a local tree for a client synchronization service

Info

Publication number: AU2021203706A1
Application number: AU2021203706A
Authority: AU
Inventors: John Lai
Original assignee: Dropbox Inc
Current assignee: Dropbox Inc
Priority date: 2017-12-28
Filing date: 2021-06-07
Publication date: 2021-07-01
Anticipated expiration: 2038-12-13
Also published as: JP7075998B2; US10733205B2; US20200349171A1; CN111448558A; JP2021509191A; US11080297B2; CA3084060C; US20230101958A1; US11188559B2; US20190205554A1; US20190205050A1; US20190205415A1; US20190205416A1; US20190205411A1; JP7068454B2; AU2018395856A1; CN111527487B; JP7212040B2; US20190205422A1; US20190205440A1

Abstract

The disclosed technology relates to a system configured to detect a first file event associated with a content item on a local file system of a client device, determine that the first file event violates a local tree constraint in a set of local tree constraints, perform a remediation associated with the violated local tree constraint, and update a local tree based on the set of observed file events, wherein the local tree represents the file system state.

Description

UPDATING A LOCAL TREE FOR A CLIENT SYNCHRONIZATION SERVICE CROSS-REFERENCE TO RELATED APPLICATIONS

[1] This application claims the priority to U.S. non-provisional application 15/858,125 filed on December 29, 2017, and to U.S. non-provisional application 15/858,146 filed on December 29, 2017, and to U.S. provisional application number 62/611,473, filed on December 28, 2017, both of which are expressly incorporated by reference herein in its entirety.

[1a] This application is related to International Application Number PCT/US2018/065352 (International Publication Number W02019/133270A1), the contents of which are incorporated herein by reference in their entirety. BACKGROUND

[2] Content management systems allow users to access and manage content items across multiple devices using a network. Some content management systems may allow users to share content items and provide additional features that aid users in collaborating using the content items. Content management systems generally store content items on servers and allow users access to the content items over a network. Some content management systems also allow for local copies to be stored on a client device in order to provide users with faster access to content items in a more natural interface (e.g., a native application or within the file system of the client device). Additionally, this allows the user to have access to the content items when the user is offline. Content management systems attempt to synchronize copies of a content item across a number of client devices and the servers so that each copy is identical. However, synchronization of content items is difficult and is associated with numerous technical obstacles.

BRIEF DESCRIPTION OF THE DRAWINGS

[3] The above-recited and other advantages and features of the present technology will become apparent by reference to specific implementations illustrated in the appended drawings. A person of ordinary skill in the art will understand that these drawings only show some examples of the present technology and would not limit the scope of the present technology to these examples. Furthermore, the skilled artisan will appreciate the principles of the present technology as described and explained with additional specificity and detail through the use of the accompanying drawings in which:

[4] FIG. 1 shows an example of a content management system and client devices, in accordance with some embodiments;

151 FIG. 2 shows an example of a client synchronization service, in accordance with some embodiments;

[61 FIG. 3 shows an example of tree data structures, in accordance with various embodiments;

[7] FIG. 4 shows an example of tree data structures, in accordance with various embodiments;

[81 FIG. 5 shows an example method for synchronizing a server state and a file system state using tree data structures, in accordance with various embodiments of the subject technology;

191 FIG. 6 shows an example method for resolving conflicts when synchronizing a server state and a file system state using tree data structures, in accordance with various embodiments of the subject technology;

[10] FIG. 7 shows an example of tree data structures illustrating a violation of a rule for an add operation, in accordance with various embodiments;

[11] FIG. 8 shows an example method for incrementally converging a server state and a file system state, in accordance with various embodiments of the subject technology;

[12] FIG. 9 shows an example of tree data structures, in accordance with various embodiments;

[13] FIG. 10 shows an example scenario;

[14] FIG. 11 shows an example Venn diagram representation of two plans of operations, in accordance with various embodiments of the subject technology.

[15] FIG. 12 shows an example method for managing changes in plans of operations, in accordance with various embodiments of the subject technology;

[16] FIG. 13 shows an example scenario, in accordance with various embodiments of the subject technology;

[17] FIG. 14 shows an example method for updating a local tree, in accordance with various embodiments of the subject technology;

[18] FIG. 15 shows an example method for updating a local tree in response to a move or rename operation, in accordance with various embodiments of the subject technology; and

[19] FIG. 16 shows an example of a system for implementing certain aspects of the present technology.

DETAILED DESCRIPTION

[20] Various examples of the present technology are discussed in detail below. While specific implementations are discussed, it should be understood that this is done for illustration purposes only. A person skilled in the relevant art will recognize that other components and configurations may be used without parting from the spirit and scope of the present technology.

[21] Various advances in computing and networking technologies have enabled content management systems to provide users with access to content items across multiple devices. The content items may include, but are not limited to, files, documents, messages (e.g., email messages or text messages), media files (e.g., photos, videos, and audio files) folders containing multiple files, or any other unit of content. Content items may be shared with multiple users, edited, deleted, added, renamed, or moved. However, synchronizing these content items across several computing devices (e.g., servers and client devices) and across several user accounts has remained flawed and rife with technological obstacles.

[22] To illustrate some of the technical obstacles, a first machine (e.g., a client device or server) may send communications to a second machine that provide information about how a user has modified content items managed by the content management system. These communications may be used by the second machine to synchronize the content items on the second machine such that actions performed on content items on the first machine are reflected in content items the second machine and the content items on the first machine are substantially identical to the content items on the second machine.

[23] However, there may be several communications sent and the communications may be received out of order as a result of various network routing protocols used by the one or more networks used to transmit the communications, the technical operations of the first or second machine, or some other reason. Furthermore, a user may be performing a large number of modifications to a large number of content items, undo previous modifications in a short amount of time, or quickly perform additional modifications to a previously modified content item or set of content items. This increases the likelihood that these communications are received out of order, certain communications are out of date, or that the second machine will perform operations on content items that are not up to date. As a result, many of the operations may not be compatible with the current state of the content items. In fact, it may be difficult to even detect whether some operations are in conflict with other operations or with the current state of the content items.

[24] Additionally, there is an inherent latency with respect to synchronization actions. For example, actions taken on the first machine is first detected by the first machine, and a communication is generated and then transmitted through a network. The communication is received by the second machine which may still be processing previous communications, processed, and actions detailed in the communications may be taken at the second machine. In this illustrative scenario, there are several points where latency is introduced by limited computing resources (e.g., bandwidth, memory, processing time, processing cycles, etc.) of the first machine, the second machine, and the network. As latency increases the likelihood that communications, for some reason, conflict with the current state of the content items are increased. Furthermore, processing these conflicted communications and resolving the conflicts also expends needless computing resources such as processing time, memory, energy, or bandwidth and further increases latency.

[25] To further complicate matters, the same or different user on the second machine and/or additional machines with access to the content items may also be performing modification to the content items. As a result, the issues above may be multiplied and additional technical issues arise as to whether local actions conflict with remote actions and/or whether local actions are operating on up to date content items.

[26] The disclosed technology addresses the need in the art for a client synchronization service for a content management system that provides a technical solution to the technical problems above as well as others. The client synchronization service may be configured to operate on a client device and identify synchronization mismatches between content items on a server of the content management system and corresponding content items on the client device. For each synchronization mismatch, the client synchronization service may identify operations needed to synchronize the content items and initiate those operations.

[27] The client synchronization service may track the status of content items on the server, the status of content items on the client device, and their synchronization state using a set of tree data structures ("trees"). According to some embodiments, a set of 3 trees may be used. The three trees may include a remote tree that represents a server state, a local tree that represents the file system state on the client device, and a sync tree that represents a merge base for the local tree and the remote tree. The merge base may be thought of as a common ancestor of the local tree and the remote tree or a last known synced state between the local tree and the remote tree. Accordingly, the client synchronization service may determine that the server state and the client device state are synchronized when all 3 trees (e.g., the remote tree, the sync tree, and the local tree) are identical.

[28] When a modification to the server state of the content items or the client device file system state ("file system state") of the content items is detected, the client synchronization service updates the appropriate tree and determines whether the server state and the file system state are synchronized based on the triumvirate of trees. Based on the update to one of the trees, the server state and the file system state may become synchronized, become unsynchronized, or become further unsynchronized. If the server state and the file system state are not synchronized, the client synchronization service may identify at least an initial set of operations needed to converge the server state and the file system state and get the server state and the file system state closer to a synchronized state.

[29] By relying on the set of tree data structures to monitor the server state and the file system state provides alternatives and/or solutions rooted in computing technology to various technical problems. For example, the client synchronization service is able to track the server state as well as the file state and store a representation of a merge base of the two states. As a result, the various embodiments of the subject technology avoid the technical problems associated with receiving a number of communications specifying how users are modifying content items remotely and determining which order these modifications should be implemented locally, whether the modifications conflict with other modifications or are out of date, and whether remote modifications conflict with local modifications performed locally by users. Many of these issues arise from other solutions not being able to track the state of the various actors involved (e.g., the server and the client device) and not being able to quickly determine whether the states are in sync. Instead, these other solutions rely on receiving instructions on how to modify content items locally, without the context of whether the server state and file system state are in sync.

[30] Furthermore, since the server state and the file system state are continuously monitored, determining whether they are synced is much more efficient in terms of procedural complexity as well as computing time and resources. As is described in further detail below, the client synchronization service enables the incremental and methodical synchronization of the server state and the file system state in a more deterministic manner. As a result, the scaling and testing of content management system features is also more efficient.

Content Management System

[31] In some embodiments, the disclosed technology is deployed in the context of a content management system having content item synchronization capabilities and collaboration features, among others. An example system configuration 100 is shown in FIG. 1A, which depicts content management system 110 interacting with client device 150.

[32] Accounts

[33] Content management system 110 can store content items in association with accounts, as well as perform a variety of content item management tasks, such as retrieve, modify, browse, and/or share the content item(s). Furthermore, content management system 110 can enable an account to access content item(s) from multiple client devices.

[34] Content management system 110 supports a plurality of accounts. An entity (user, group of users, team, company, etc.) can create an account with content management system, and account details can be stored in account database 140. Account database 140 can store profile information for registered entities. In some cases, profile information for registered entities includes a username and/or email address. Account database 140 can include account management information, such as account type (e.g. various tiers of free or paid accounts), storage space allocated, storage space used, client devices 150 having a registered content management client application 152 resident thereon, security settings, personal configuration settings, etc.

[35] Account database 140 can store groups of accounts associated with an entity. Groups can have permissions based on group policies and/or access control lists, and members of the groups can inherit the permissions. For example, a marketing group can have access to one set of content items while an engineering group can have access to another set of content items. An administrator group can modify groups, modify user accounts, etc.

[36] Content Item Storage

[37] A feature of content management system 110 is the storage of content items, which can be stored in content storage 142. Content items can be any digital data such as documents, collaboration content items, text files, audio files, image files, video files, webpages, executable files, binary files, etc. A content item can also include collections or other mechanisms for grouping content items together with different behaviors, such as folders, zip files, playlists, albums, etc. A collection can refer to a folder, or a plurality of content items that are related or grouped by a common attribute. In some embodiments, content storage 142 is combined with other types of storage or databases to handle specific functions. Content storage 142 can store content items, while metadata regarding the content items can be stored in metadata database 146. Likewise, data regarding where a content item is stored in content storage 142 can be stored in content directory 144. Additionally, data regarding changes, access, etc. can be stored in server filejournal 148. Each of the various storages/databases such as content storage 142, content directory 144, server filejournal 148, and metadata database 146 can be comprised of more than one such storage or database and can be distributed over many devices and locations. Other configurations are also possible. For example, data from content storage 142, content directory 144, server filejournal 148, and/or metadata database 146 may be combined into one or more content storages or databases or further segmented into additional content storages or databases. Thus, content management system 110 may include more or less storages and/or databases than shown in FIG. 1.

[38] In some embodiments, content storage 142 is associated with at least one content storage service 116, which includes software or other processor executable instructions for managing the storage of content items including, but not limited to, receiving content items for storage, preparing content items for storage, selecting a storage location for the content item, retrieving content items from storage, etc. In some embodiments, content storage service 116 can divide a content item into smaller chunks for storage at content storage 142. The location of each chunk making up a content item can be recorded in content directory 144. Content directory 144 can include a content entry for each content item stored in content storage 142. The content entry can be associated with a unique ID, which identifies a content item.

[39] In some embodiments, the unique ID, which identifies a content item in content directory 144, can be derived from a deterministic hash function. This method of deriving a unique ID for a content item can ensure that content item duplicates are recognized as such since the deterministic hash function will output the same identifier for every copy of the same content item, but will output a different identifier for a different content item. Using this methodology, content storage service 116 can output a unique ID for each content item.

[40] Content storage service 116 can also designate or record a content path for a content item in metadata database 146. The content path can include the name of the content item and/or folder hierarchy associated with the content item. For example, the content path can include a folder or path of folders in which the content item is stored in a local file system on a client device. While content items are stored in content storage 142 in blocks and may not be stored under a tree like directory structure, such directory structure is a comfortable navigation structure for users. Content storage service 116 can define or record a content path for a content item wherein the "root" node of a directory structure can be a namespace for each account. Within the namespace can be a directory structure defined by a user of an account and/or content storage service 116. Metadata database 146 can store the content path for each content item as part of a content entry.

[41] In some embodiments the namespace can include additional namespaces nested in the directory structure as if they are stored within the root node. This can occur when an account has access to a shared collection. Shared collections can be assigned their own namespace within content management system 110. While some shared collections are actually a root node for the shared collection, they are located subordinate to the account namespace in the directory structure, and can appear as a folder within a folder for the account. As addressed above, the directory structure is merely a comfortable navigation structure for users, but does not correlate to storage locations of content items in content storage 142.

[42] While the directory structure in which an account views content items does not correlate to storage locations at content management system 110, the directory structure can correlate to storage locations on client device 150 depending on the file system used by client device 150.

[43] As addressed above, a content entry in content directory 144 can also include the location of each chunk making up a content item. More specifically, the content entry can include content pointers that identify the location in content storage 142 of the chunks that make up the content item.

[44] In addition to a content path and content pointer, a content entry in content directory 144 can also include a user account identifier that identifies the user account that has access to the content item and/or a group identifier that identifies a group with access to the content item and/or a namespace to which the content entry belongs.

[45] Content storage service 116 can decrease the amount of storage space required by identifying duplicate content items or duplicate blocks that make up a content item or versions of a content item. Instead of storing multiple copies, content storage 142 can store a single copy of the content item or block of the content item and content directory 144 can include a pointer or other mechanism to link the duplicates to the single copy.

[46] Content storage service 116 can also store metadata describing content items, content item types, folders, file path, and/or the relationship of content items to various accounts, collections, or groups in metadata database 146, in association with the unique ID of the content item.

[47] Content storage service 116 can also store a log of data regarding changes, access, etc. in server file journal 148. Server file journal 148 can include the unique ID of the content item and a description of the change or access action along with a time stamp or version number and any other relevant data. Server filejournal 148 can also include pointers to blocks affected by the change or content item access. Content storage service can provide the ability to undo operations, by using a content item version control that tracks changes to content items, different versions of content items (including diverging version trees), and a change history that can be acquired from the server filejournal 148.

[48] Content Item Synchronization

[49] Another feature of content management system 110 is synchronization of content items with at least one client device 150. Client device(s) can take different forms and have different capabilities. For example, client device 150t is a computing device having a local file system accessible by multiple applications resident thereon. Client device 1502 is a computing device wherein content items are only accessible to a specific application or by permission given by the specific application, and the content items are typically stored either in an application specific space or in the cloud. Client device 1503 is any client device accessing content management system 110 via a web browser and accessing content items via a web interface. While example client devices 150i, 1502, and 1503 are depicted in form factors such as a laptop, mobile device, or web browser, it should be understood that the descriptions thereof are not limited to devices of these example form factors. For example a mobile device such as client 1502 might have a local file system accessible by multiple applications resident thereon, or client 150 2 might access content management system 110 via a web browser. As such, the form factor should not be considered limiting when considering client 150's capabilities. One or more functions described herein with respect to client device 150 may or may not be available on every client device depending on the specific capabilities of the device - the file access model being one such capability.

[50] In many embodiments, client devices are associated with an account of content management system 110, but in some embodiments client devices can access content using shared links and do not require an account.

[51] As noted above, some client devices can access content management system 110 using a web browser. However, client devices can also access content management system 110 using client application 152 stored and running on client device 150. Client application 152 can include a client synchronization service 156.

[52] Client synchronization service 156 can be in communication with server synchronization service 112 to synchronize changes to content items between client device 150 and content management system 110.

[53] Client device 150 can synchronize content with content management system 110 via client synchronization service 156. The synchronization can be platform agnostic. That is, content can be synchronized across multiple client devices of varying type, capabilities, operating systems, etc. Client synchronization service 156 can synchronize any changes (new, deleted, modified, copied, or moved content items) to content items in a designated location of a file system of client device 150.

[54] Content items can be synchronized from client device 150 to content management system 110, and vice versa. In embodiments wherein synchronization is from client device 150 to content management system 110, a user can manipulate content items directly from the file system of client device 150, while client synchronization service 156 can monitor directory on client device 150 for changes to files within the monitored folders.

[55] When client synchronization service 156 detects a write, move, copy, or delete of content in a directory that it monitors, client synchronization service 156 can synchronize the changes to content management system service 116. In some embodiments, client synchronization service 156 can perform some functions of content management system service 116 including functions addressed above such as dividing the content item into blocks, hashing the content item to generate a unique identifier, etc. Client synchronization service 156 can index content within client storage index 164 and save the result in storage index 164. Indexing can include storing paths plus a unique server identifier, and a unique client identifier for each content item. In some embodiments, client synchronization service 156 learns the unique server identifier from server synchronization service 112, and learns the unique client identifier from the operating system of client device 150. ,

[56] Client synchronization service 156 can use storage index 164 to facilitate the synchronization of at least a portion of the content within client storage with content associated with a user account on content management system 110. For example, client synchronization service 156 can compare storage index 164 with content management system 110 and detect differences between content on client storage and content associated with a user account on content management system 110. Client synchronization service 156 can then attempt to reconcile differences by uploading, downloading, modifying, and deleting content on client storage as appropriate. Content storage service 116 can store the changed or new block for the content item and update server file journal 148, metadata database 146, content directory 144, content storage 142, account database 140, etc. as appropriate.

[57] When synchronizing from content management system 110 to client device 150, a mount, modification, addition, deletion, move of a content item recorded in server file journal 148 can trigger a notification to be sent to client device 150 using notification service 117. When client device 150 is informed of the change arequest changes listed in server file journal 148 since the last synchronization point known to the client device. When client device 150 determines that it is out of synchronization with content management system 110, client synchronization service 156 requests content item blocks including the changes, and updates its local copy of the changed content items.

[58] In some embodiments, storage index 164 stores tree data structures wherein one tree reflects the latest representation of a directory according to server synchronization service 112, while another tree reflects the latest representation of the directory according to client synchronization service 156. Client synchronization service can work to ensure that the tree structures match by requesting data from server synchronization service 112 or committing changes on client device 150 to content management system 110.

[59] Sometimes client device 150 might not have a network connection available. In this scenario, client synchronization service 156 can monitor the linked collection for content item changes and queue those changes for later synchronization to content management system 110 when a network connection is available. Similarly, a user can manually start, stop, pause, or resume synchronization with content management system 110.

[60] Client synchronization service 156 can synchronize all content associated with a particular user account on content management system 110. Alternatively, client synchronization service 156 can selectively synchronize a portion of the content of the total content associated with the particular user account on content management system 110. Selectively synchronizing only a portion of the content can preserve space on client device 150 and save bandwidth.

[61] In some embodiments, client synchronization service 156 selectively stores a portion of the content associated with the particular user account and stores placeholder content items in client storage for the remainder portion of the content. For example, client synchronization service 156 can store a placeholder content item that has the same filename, path, extension, metadata, of its respective complete content item on content management system 110, but lacking the data of the complete content item. The placeholder content item can be a few bytes or less in size while the respective complete content item might be significantly larger. After client device 150 attempts to access the content item, client synchronization service 156 can retrieve the data of the content item from content management system 110 and provide the complete content item to accessing client device 150. This approach can provide significant space and bandwidth savings while still providing full access to a user's content on content management system 110.

[62] Collaboration features

[63] Another feature of content management system 110 is to facilitate collaboration between users. Collaboration features include content item sharing, commenting on content items, co-working on content items, instant messaging, providing presence and seen state information regarding content items, etc.

[64] Sharing

[65] Content management system 110 can manage sharing content via sharing service 128. Sharing content by providing a link to the content can include making the content item accessible from any computing device in network communication with content management system 110. However, in some embodiments a link can be associated with access restrictions enforced by content management system 110 and access control list 145. Sharing content can also include linking content using sharing service 128 to share content within content management system 110 with at least one additional user account (in addition to the original user account associated with the content item) so that each user account has access to the content item. The additional user account can gain access to the content by accepting the content, which will then be accessible through either web interface service 124 or directly from within the directory structure associated with their account on client device 150. The sharing can be performed in a platform agnostic manner. That is, the content can be shared across multiple client devices 150 of varying type, capabilities, operating systems, etc. The content can also be shared across varying types of user accounts.

[66] To share a content item within content management system 110 sharing service 128 can add a user account identifier or multiple user account identifiers to a content entry in access control list database 145 associated with the content item, thus granting the added user account access to the content item. Sharing service 128 can also remove user account identifiers from a content entry to restrict a user account's access to the content item. Sharing service 128 can record content item identifiers, user account identifiers given access to a content item, and access levels in access control list database 145. For example, in some embodiments, user account identifiers associated with a single content entry can specify different permissions for respective user account identifiers with respect to the associated content item.

[671 To share content items outside of content management system 110, sharing service 128 can generate a custom network address, such as a uniform resource locator (URL), which allows any web browser to access the content item or collection in content management system 110 without any authentication. To accomplish this, sharing service 128 can include content identification data in the generated URL, which can later be used to properly identify and return the requested content item. For example, sharing service 128 can include the account identifier and the content path or a content item identifying code in the generated URL. Upon selection of the URL, the content identification data included in the URL can be transmitted to content management system 110, which can use the received content identification data to identify the appropriate content item and return the content item.

[68] In addition to generating the URL, sharing service 128 can also be configured to record in access control list database 145 that a URL to the content item has been created. In some embodiments, the content entry associated with a content item can include a URL flag indicating whether a URL to the content item has been created. For example, the URL flag can be a Boolean value initially set to 0 or false to indicate that a URL to the content item has not been created. Sharing service 128 can change the value of the flag to 1 or true after generating a URL to the content item.

[69] In some embodiments, sharing service 128 can associate a set of permissions to a URL for a content item. For example, if a user attempts to access the content item via the URL, sharing service 128 can provide a limited set of permissions for the content item. Examples of limited permissions include restrictions that the user cannot download the content item, save the content item, copy the content item, modify the content item, etc. In some embodiments, limited permissions include restrictions that only permit a content item to be accessed from with a specified domain, i.e., from within a corporate network domain, or by accounts associated with a specified domain, e.g., accounts associated with a company account (e.g., @acme.com).

[70] In some embodiments, sharing service 128 can also be configured to deactivate a generated URL. For example, each content entry can also include a URL active flag indicating whether the content should be returned in response to a request from the generated URL. For example, sharing service 128 can only return a content item requested by a generated link if the URL active flag is set to 1 or true. Thus, access to a content item for which a URL has been generated can be easily restricted by changing the value of the URL active flag. This allows a user to restrict access to the shared content item without having to move the content item or delete the generated URL. Likewise, sharing service 128 can reactivate the URL by again changing the value of the URL active flag to 1 or true. A user can thus easily restore access to the content item without the need to generate a new URL.

[71] In some embodiments, content management system 110 can designate a URL for uploading a content item. For example, a first user with a user account can request such a URL, provide the URL to a contributing user and the contributing user can upload a content item to the first user's user account using the URL.

[72] Team Service

[73] In some embodiments content management system 110 includes team service 130. Team service 130 can provide functionality for creating and managing defined teams of user accounts. Teams can be created for a company, with sub-teams (e.g., business units, or project teams, etc.), and user accounts assigned to teams and sub teams, or teams can be created for any defined group of user accounts. Team's service 130 can provide a common shared space for the team, private user account folders, and access limited shared folders. Team's service can also provide a management interface for an administrator to manage collections and content items within team, and can manage user accounts that are associated with the team.

[74] Authorization Service

[75] In some embodiments, content management system 110 includes authorization service 132. Authorization service 132 ensures that a user account attempting to access a namespace has appropriate rights to access the namespace. Authorization service 132 can receive a token from client application 152 that follows a request to access a namespace and can return the capabilities permitted to the user account. For user accounts with multiple levels of access (e.g. a user account with user rights and administrator rights) authorization service 132 can also require explicit privilege escalation to avoid unintentional actions by administrators

[76] Presence and Seen State

[77] In some embodiments, content management system can provide information about how users with which a content item is shared are interacting or have interacted with the content item. In some embodiments, content management system 110 can report that a user with which a content item is shared is currently viewing the content item. For example, client collaboration service 160 can notify notifications service 117 when client device 150 is accessing the content item. Notifications service 117 can then notify all client devices of other users having access to the same content item of the presence of the user of client device 150 with respect to the content item.

[78] In some embodiments, content management system 110 can report a history of user interaction with a shared content item. Collaboration service 126 can query data sources such as metadata database 146 and server filejournal 148 to determine that a user has saved the content item, that a user has yet to view the content item, etc., and disseminate this status information using notification service 117 to other users so that they can know who currently is or has viewed or modified the content item.

[79] Collaboration service 126 can facilitate comments associated with content, even if a content item does not natively support commenting functionality. Such comments can be stored in metadata database 146.

[80] Collaboration service 126 can originate and transmit notifications for users. For example, a user can mention another user in a comment and collaboration service

126 can send a notification to that user that he has been mentioned in the comment. Various other content item events can trigger notifications, including deleting a content item, sharing a content item, etc.

[81] Collaboration service 126 can provide a messaging platform whereby users can send and receive instant messages, voice calls, emails, etc.

[82] Collaboration Content Items

[83] In some embodiments content management service can also include Collaborative document service 134 which can provide an interactive content item collaboration platform whereby users can simultaneously create collaboration content items, comment in the collaboration content items, and manage tasks within the collaboration content items. Collaboration content items can be files that users can create and edit using a collaboration content item editor, and can contain collaboration content item elements. Collaboration content item elements may include a collaboration content item identifier, one or more author identifiers, collaboration content item text, collaboration content item attributes, interaction information, comments, sharing users, etc. Collaboration content item elements can be stored as database entities, which allows for searching and retrieving the collaboration content items. Multiple users may access, view, edit, and collaborate on collaboration content items at the same time or at different times. In some embodiments this can be managed by requiring two users access a content item through a web interface and there they can work on the same copy of the content item at the same time.

[84] Collaboration Companion Interface

[85] In some embodiments client collaboration service 160 can provide a native application companion interface for the purpose of displaying information relevant to a content item being presented on client device 150. In embodiments wherein a content item is accessed by a native application stored and executed on client device 150, where the content item is in a designated location of the file system of client device 150 such that the content item is managed by content application 152, the native application may not provide any native way to display the above addressed collaboration data. In such embodiments, client collaboration service 160 can detect that a user has opened a content item, and can provide an overlay with additional information for the content item, such as collaboration data. For example, the additional information can include comments for the content item, status of the content item, activity of other users previously or currently viewing the content item. Such an overlay can warn a user that changes might be lost because another user is currently editing the content item.

[86] In some embodiments, one or more of the services or storages/databases discussed above can be accessed using public or private application programming interfaces.

[87] Certain software applications can access content storage 142 via an API on behalf of a user. For example, a software package such as an application running on client device 150, can programmatically make API calls directly to content management system 110 when a user provides authentication credentials, to read, write, create, delete, share, or otherwise manipulate content.

[88] A user can view or manipulate content stored in a user account via a web interface generated and served by web interface service 124. For example, the user can navigate in a web browser to a web address provided by content management system 110. Changes or updates to content in the content storage 142 made through the web interface, such as uploading a new version of a content item, can be propagated back to other client devices associated with the user's account. For example, multiple client devices, each with their own client software, can be associated with a single account and content items in the account can be synchronized between each of the multiple client devices.

[89] Client device 150 can connect to content management system 110 on behalf of a user. A user can directly interact with client device 150, for example when client device 150 is a desktop or laptop computer, phone, television, internet-of-things device, etc. Alternatively or additionally, client device 150 can act on behalf of the user without the user having physical access to client device 150, for example when client device 150 is a server.

[90] Some features of client device 150 are enabled by an application installed on client device 150. In some embodiments, the application can include a content management system specific component. For example, the content management system specific component can be a stand-alone application 152, one or more application plug-ins, and/or a browser extension. However, the user can also interact with content management system 110 via a third-party application, such as a web browser, that resides on client device 150 and is configured to communicate with content management system 110. In various implementations, the client-side application 152 can present a user interface (UI) for a user to interact with content management system 110. For example, the user can interact with the content management system 110 via a file system explorer integrated with the file system or via a webpage displayed using a web browser application.

[91] In some embodiments, client application 152 can be configured to manage and synchronize content for more than one account of content management system 110. In such embodiments client application 152 can remain logged into multiple accounts and provide normal services for the multiple accounts. In some embodiments, each account can appear as folder in a file system, and all content items within that folder can be synchronized with content management system 110. In some embodiments, client application 152 can include a selector to choose one of the multiple accounts to be the primary account or default account.

[92] While content management system 110 is presented with specific components, it should be understood by one skilled in the art, that the architectural configuration of system 100 is simply one possible configuration and that other configurations with more or fewer components are possible. Further, a service can have more or less functionality, even including functionality described as being with another service. Moreover, features described herein with respect to an embodiment can be combined with features described with respect to another embodiment.

[93] While system 100 is presented with specific components, it should be understood by one skilled in the art, that the architectural configuration of system 100 is simply one possible configuration and that other configurations with more or fewer components are possible.

Client Synchronization Service

[94] FIG. 2 shows an example of a client synchronization service 156, in accordance with some embodiments. According to some embodiments, client synchronization service 156 may be implemented in the client device of FIG. 1. However, in other embodiments, client synchronization service 156 may be implemented on another computing device. Client synchronization service 156 is configured to synchronize changes to content items between a content management system and the client device on which client synchronization service 156 runs.

[95] Client synchronization service 156 may include file system interface 205, server interface 210, tree storage 220, planner 225, and scheduler 230. Additional or alternative components may also be included. High level descriptions of client synchronization service 156 and its components are discussed below with respect to FIG. 2. However, further details and embodiments of client synchronization service 156 and its components are discussed throughout.

[96] File system interface 205 is configured to process changes to content items on the local filesystem of the client device and update the local tree. For example, file system interface 205 can be in communication with client synchronization service 156 of FIG. 1 detect changes to content items on the local filesystem of the client device. Changes may also be made and detected via client application 152 of FIG. 1. File system interface 205 may make updates to the local tree may be made based on the changes (new, deleted, modified, copied, renamed, or moved content items) to content items on the client device.

[97] Server interface 210 is configured to aid in the processing of remote changes to content items at a remote storage of the content management system and updating of the remote tree. For example, server interface 210 can be in communication with server synchronization service 112 of FIG. 1 to synchronize changes to content items between client device 150 and content management system 110. Changes (new, deleted, modified, copied, renamed, or moved content items) to content items at content management system 110 may be detected and updates may be made to the remote tree to reflect the changes at content management system 110.

[98] Tree storage 220 is configured to store and maintain the tree data structures used by client synchronization service 156. For example, tree storage 220 may store the local tree, the sync tree, and the remote tree. According to some embodiments, tree storage 200 may store the tree data structures in persistent memory (e.g., a hard disk or other secondary storage device) as well as in main memory (e.g., RAM or other primary storage device) in order to reduce latency and response time. For example, on start-up of the client device or client synchronization service 156, the tree data structures may be retrieved from persistent memory and loaded into main memory. Tree storage 220 may access and update the tree data structures on main memory and, before the client device or client synchronization service 156 is shut down, tree storage 220 may store the updated tree data structures on persistent memory. Because main memory is expensive in cost and often limited in size on most client devices, additional technological improvements are implemented to decrease the footprint of the tree data structures on main memory. These technological solutions are described further below.

[99] Planner 225 is configured to detect differences between the server state associated with the content management system and the file system state associated with the client device based on the state of the tree data structures. For example, planner 225 may determine if there is a difference between the remote tree and the sync tree. A difference between the remote tree and the sync tree indicates that an action performed remotely on one or more content items stored at the content management system has caused the server state and the file system state to become out of sync. Similarly, planner 225 may also determine if there is a difference between the local tree and the sync tree. A difference between the local tree and the sync tree indicates that an action performed locally on one or more content items stored on the client device has caused the server state and the file system state to become out of sync. If a difference is detected, planner 225 generates a set of operations that synchronize the tree data structures.

11001 In some scenarios, a set of operations generated based on a difference between the remote tree and the sync tree and a set of operations generated based on a difference between the local tree and the sync tree may conflict. Planner 225 is may also be configured to merge the two sets of operations into a single merged plan of operations.

11011 Scheduler 230 is configured to take the generated plan of operations and manage the execution of those operations. According to some embodiments, scheduler 230 converts each operation in the plan of operations into a series of one or more tasks that need to be executed in order to perform the operation. In some scenarios, some tasks may become out dated or no longer relevant. Scheduler 230 is configured to identify those tasks and cancel them.

Tree Data Structures

[102] FIG. 3 shows an example of tree data structures, in accordance with various embodiments. The tree data structures may be stored at the client device and managed by a client synchronization service such as client synchronization service 156 in FIG. 2. In FIG. 3, the tree data structures are shown including remote tree 310, sync tree 330, andlocalwee 350.

11031 Remote tree 310 represents a server state or the state of content items stored remotely from the client device (e.g., on a server of the content management system). Local tree 350 represents a file system state or the state of the corresponding content items stored locally on the client device. Sync tree 330 represents a merge base for the local tree and the remote tree. The merge base may be thought of as a common ancestor of the local tree and the remote tree or a last known synced state between the local tree and the remote tree.

11041 Each tree data structure (e.g., remote tree 310, sync tree 330, or local tree 350) may include one or more nodes. Each node may have one or more child nodes and the parent-child relationship is represented by an edge. For example, remote tree 310 includes nodes 312 and 314. Node 312 is a parent of node 314 and node 314 is a child ofnode 312. This parent-child relationship is represented by edge 316. A root node, such as root node 312, does not have a parent node. A leaf node, such as node 314, does not have a child node.

11051 Each node in a tree data structure may represent a content item (e.g., a file, document, folder, etc.). For example, root node 312 may represent the root folder associated with the content management system and node 314 may represent a file (e.g., a text file named "Foo.txt") located in that root folder. Each node in a tree data structure may contain data such as, for example, a directory file identifier ("DirFilelD") specifying the file identifier of a parent node of the content item, a file name for the content item, a file identifier for the content item, and metadata for the content item. In some embodiments each node in a tree data structure may be keyed or referenced by its file identifier and have a unique path from the root to the node.

11061 As described above, a client synchronization service may determine that the server state and the file system state of the client device are synchronized when all 3 trees (e.g., remote tree 310, sync tree 330, and local tree 350) are identical. In other words, the trees are in sync when their tree structures and the relationships that they express are identical and the data contained in their nodes are identical as well. Conversely, the trees are not in sync if the 3 trees are not identical. In the example scenario illustrated in FIG. 3, remote tree 310, sync tree 330, and local tree 350 are shown as being identical and in sync and, as a result, the server state and the file system state are synchronized.

Tracking Changes Using Tree Data Structures

[107] FIG. 4 shows an example of tree data structures, in accordance with various embodiments. As with the tree data structures shown in FIG. 3, the tree data structures shown in FIG. 4 (including remote tree 410, sync tree 430, and local tree 450) may be stored at the client device and managed by a client synchronization service such as client synchronization service 156 in FIG. 2. In FIG. 3, the tree data structures are shown.

[108] FIG. 4 shows a scenario after a previously synchronized state, such as the scenario illustrated in FIG. 3, additional actions are performed on the content items represented in the trees to modify the content items such that the trees are no longer in sync. Sync tree 430 maintains a representation of the previously known synchronized state and may be used by the client synchronization service to identify the differences between the server state and the file system state as well as generate operations for the content management system and/or the client device to perform to converge so that the server state and the file system state are synchronized.

[109] For example, a user (the same user as the user associated with the client device or a different user with access to the content item) may make modifications to the "foo.txt" content item stored by the content management system. This content item is represented by node 414 in remote tree 410. The modification shown in the remote tree 410 is aremoval (e.g., aremoval of the content item from a space managed bythe content management system) or delete of the foo.txt content item. These modifications may be performed, for example, on another client device and the modifications were synced to the content item stored by the content management system or content item stored by the content management system via a web browser.

11101 When the change is made on the content management system, the content management system generates modification data specifying the change made and transmits the modification data to the client synchronization service on the client device. The client synchronization service updates the remote tree representing the server state for the content items stored by the content management system based on the modification data. For example, in remote tree 410, node 414 representing the foo.txt content item is shown as deleted.

[111] The client synchronization service may identify a difference between remote tree 410 and sync tree 430 and, as a result, determine that a modification of the content items at the content management system has caused the server state and the file system state to no longer be in sync. The client synchronization service may further generate and execute a set or sequence of operations for the content items stored on the client device that are configured to converge the server state and the file system state so that they will be in sync.

11121 Additionally or alternatively, a user (the same user as the user associated with modifications at the content management system or a different user with access to the content item) may make modifications to the content items stored locally on the client device that are associated with the content management system. For example, the user may add a folder "/bar" to the "/root" folder and add a "Hi.doc" document to the "/bar" folder.

11131 When the change is made on the client device, the client device (e.g., client synchronization service 156 or client application 152 of FIG. 1) generates modification data specifying the change made and passes the modification data to the client synchronization service on the client device. The client synchronization service updates the local tree representing the file system state for the content items stored on the client device based on the modification data. For example, in local tree 450, node 452 and node 454 are shown as added. Node 452 and node 454 represent the "/bar" folder and the "Hi.doc" document respectively.

11141 The client synchronization service may identify a difference between local tree 450 and sync tree 430 and, as a result, determine that a modification of the content items at the client device has caused the server state and the file system state to no longer be in sync. The client synchronization service may further generate a set or sequence of operations for the content items stored by the content management system that are configured to converge the server state and the file system state so that they will be in sync. These operations may be transmitted to the content management system for execution.

11151 As seen in FIG. 4, modifications to content items stored on the client device and content items stored by the content management system may occur at substantially the same time or within a particular time period. These modifications can be reflected in the tree data structures and used by the client synchronization service to generate operations for the client device and for the content management system in parallel. In other scenarios, however, modifications may not necessarily occur within the same time period and operations may be generated in an as-needed manner. Furthermore, although FIG. 4 illustrates scenarios for adding content items and deleting content items, other types of modifications such as, editing, renaming, copying, or moving content items are also supported.

11161 According to various embodiments, identifying a difference between two tree data structures and generating operations may involve checking each node in both tree data structures and determining whether an action has been performed on the node. The actions may include, for example, the addition of the node, the deletion of the node, the editing of the node, or the moving of the node. These actions may then be used to generate the operations configured to converge the server state and the file system state.

11171 For example, if the two tree data structures are a sync tree and a remote tree, the client synchronization service may identify each node in the sync tree by, for example, requesting the file identifiers of all nodes in the sync tree. For each node or file identifier for the node in the sync tree, the client synchronization service may determine if the node or file identifier is also in the remote tree. A node or file identifier in the sync tree that is not found in the remote tree may indicate that the node has been deleted from the server state that is represented by the remote tree. Accordingly, the client synchronization service may determine that a delete action has occurred on the remote tree. If the node or file identifier for the node is found in the remote tree, the client synchronization service may check whether the node in the remote tree has been edited or moved.

11181 To determine whether the node in the remote tree has been edited with respect to the node in the sync tree, the client synchronization service may compare the metadata for the node in the sync tree with the metadata for the corresponding node (e.g., the node with the same file identifier) in the remote tree. The metadata may include information that may be used to determine whether the content item represented by the node has been edited. For example, the metadata may include one or more hash values that are generated based on the data in the content item or a portion thereof. The metadata may additionally or alternatively include a size value, a last modified value, or other value for the content item. The metadata for the node in the sync tree may be compared with the metadata for the node in the remote tree. If the metadata do not match, an edit of the content item may have been edited in the server state represented by the remote tree. Accordingly, the client synchronization service may determine that an edit action has occurred for the node on the remote tree. If the metadata matches, no edit may have occurred.

[119] To determine whether the node in the remote tree has been moved, the client synchronization service may compare the location for the node in the sync tree with the location for the corresponding node (e.g., the node with the same file identifier) in the remote tree. The location may include, for example, a path where the node is located, a file name, and/or a directory file identifier ("DirFilelD") specifying the file identifier of the node's parent. If the locations match, no move may have occurred. On the other hand, if the locations do not match, a move of the content item may have occurred in the server state represented by the remote tree. Accordingly, the client synchronization service may determine that a move action has occurred for the node on the remote tree.

[120] To determine whether a node has been added to the remote tree, the client synchronization service may identify any nodes or file identifiers in the remote tree that are not found in the sync tree. If a node orfile identifier is found in the remote tree and not found in the sync tree, the client synchronization service may determine that an add action of this node has occurred on the remote tree representing the server state.

[121] Although the example above is described with respect to the sync tree and the remote tree, in other embodiments, a similar process may occur with the sync tree and a local tree in order to identify a difference between the sync tree and the local tree and determine which actions have occurred on the local tree representing the file system state.

Synchronization Using Tree Data Structures

[122] FIG. 5 shows an example method for synchronizing a server state and a file system state using tree data structures, in accordance with various embodiments of the subject technology. Although the methods and processes described herein may be shown with certain steps and operations in a particular order, additional, fewer, or alternative steps and operations performed in similar or alternative orders, or in parallel, are within the scope of various embodiments unless otherwise stated. The method 500 may be implemented by a system such as, for example, client synchronization service 156 of FIG. 2, running on a client device.

11231 The system is configured to identify a difference between a remote tree representing a server state for content items stored by the content management system, a local tree representing the file system state for the corresponding content items stored on the client device, and a sync tree representing a known synced state between the server state and the file system state. Based on these differences, a set of operations may be generated that, if executed, are configured to converge the server state and the file system state towards a synchronized state where the three tree data structures would be identical.

11241 For example, at operation 505, the system may receive modification data for content items stored by a content management system or on a client device. The modification data may be used to update a remote tree or a local tree at operation 510.

11251 The modification data is specifies what changes are done to one or more content items associated with a content management service. Accordingly, the modification data may be received from the content management system or from the client device (e.g., from client application 152 running on client device 150 in FIG. 1). Modification data received from the content management system may be referred to as server modification data. Server modification data specifies what changes are done to one or more content items by the content management system and may be used to update the remote tree at operation 510. Modification data received from the client device may be referred to as client modification data. Client modification data specifies what changes are done to one or more content items on the client device and may be used to update the local tree at operation 510.

11261 At operation 515, the system may determine whether a server state for content items stored by the content management system and a file system state for the content items stored on the client device are in sync. Because the local tree and the remote tree are representative of the file system state and the server state and are continually being updated to track changes that occur at the content management system and the client device, determining whether the server state and the file system state are in sync may be done by comparing the local tree and/or the remote tree to the sync tree to find differences between the trees. This process of finding differences between the trees is sometimes referred to as "diffing" the trees.

11271 According to some embodiments and scenarios, determining whether the server state and the file system state are in sync may include one or more of identifying differences between the remote tree and the sync tree and/or identifying differences between the local tree and the sync tree. Differences between the remote tree and sync tree may indicate the occurrence of changes to content items stored by the content management system that may not be reflected at the client device. Similarly, differences between the local tree and sync tree may indicate the occurrence of changes to content items stored at the client device that may not be reflected at the content management system.

11281 If there are no differences between the trees, the server state and the file system state are in sync and no synchronization actions are needed. Accordingly, the method may return to operation 505 and await new modification data. On the other hand, if differences are detected, the system may generate a set of operations configured to converge the server state and the file system state at operation 520.

11291 The set of operations generated depends on the one or more differences that are detected. For example, if the difference between two trees is an added content item, the generated set of operations may include retrieving the added content item and adding it. If the difference between two trees is a deletion of a content item, the generated set of operations may include deleting the content item. According to some embodiments, the set of operations may also include a number of checks to ensure tree constraints are maintained. As will be described further below, the set of operations may conflict with the current state of the server state, the file system state, or other operations that are pending execution. Accordingly, the system may also resolve these conflicts before proceeding.

11301 As noted above, if there are differences between the remote tree and sync tree, changes to content items stored by the content management system may have occurred that may not be reflected at the client device. Accordingly, in this scenario, the system may generate a client set of operations configured to operate on the content items stored on the client device to converge the server state and the file system state and this client set of operations may be provided to the client device for execution at operation 525.

11311 On the other hand, if there are differences between the local tree and sync tree, changes to content items stored at the client device may have occurred that may not be reflected at the content management system. Accordingly, in this scenario, the system may generate a server set of operations configured to operate on the content items stored by the content management system to converge the server state and the file system state and this server set of operations may be provided to the content management system for execution at operation 525. In some cases, both cases may be true and a client set of operations and a server set of operations may be generated and provided to their intended recipients at operation 525.

11321 Once the set(s) of operations are provided to the intended recipient(s), the method may return to operation 505 and await new modification data. The set(s) of operations may provide one or more steps towards the convergence of the server state and the file system state or provide all steps needed to sync the server state and the file system state. For example, the content management system may receive the server set of operations and execute the server set of operations on content items stored by the content management system. This execution of the server set of operations causes changes to the content items stored by the content management system, which are detected and specified in server modification data, which is transmitted back to the system. The system may then update the remote tree and determine whether the server state and the file system state are in sync.

11331 The client device may receive the client set of operations and execute the client set of operations on content items stored on the client device. This execution of the client set of operations causes changes to the content items stored on the client device, which are detected and specified in client modification data, which is passed to the system. The system may then update the local tree and determine whether the server state and the file system state are in sync. These operations of method 500 may continue until the server state and the file system state are in sync.

11341 The operations of method 500 are described with respect to a client side and a server side (e.g., a local tree and a remote tree, a file system state and a server state, a client set of operations and a server set of operations, client modification data and server modification data). In various embodiments the operations associated with the two sides may occur in parallel, in sequence, in isolation of the other side, or a combination.

11351 As will be discussed in further detail, in accordance with some embodiments, before the operations are provided for execution, the system may check the operations to determine whether they comply with a set of rules or invariants. If an operation violates a rule, the system executes a resolution process associated with the violation of the rule.

11361 Additionally, in accordance with some embodiments, the system (e.g., scheduler 230 of client synchronization service 156 in FIG. 2) may manage the execution of the set of operations. For example, each operation in the set of operations may be associated with a task, an execution thread, series of steps, or instructions. The system may be configured to execute the task, thread, step, or instructions and interface with the client device and/or the content management system to execute the set of operations and converge the server state and the file system state.

Conflict Handling

11371 As described above with respect to FIG. 5, differences between a sync tree and a remote tree are identified and used to generate a client set of operations configured to converge the server state and the file system state. However, in some cases, the client set of operations may conflict with the current state of a local tree. Similarly, differences between the sync tree and the local tree are identified and used to generate a server set of operations configured to converge the server state and the file system state. However, the server set of operations may conflict with the current state of the remote tree. Additionally or alternatively, the client set of operations and the server set of operations may conflict with one another or violate another rule or invariant maintained by the system. Accordingly, various embodiments of the subject technology provide additional technical improvements by resolving these conflicts.

11381 For example, planner 225 in client synchronization service 156 of FIG. 2 may identify an operation in a set of operations (e.g., the client set of operations or the server set of operations) that conflicts with a rule. Each rule used to identify a conflict may also be associated with a resolution for the conflict. The client synchronization service may update the set of operations based on the resolution for the conflict or perform resolve the conflict by performing operations associated with the resolutions for the conflict before providing the set of operations for execution.

11391 FIG. 6 shows an example method 600 for resolving conflicts when synchronizing a server state and a file system state using tree data structures, in accordance with various embodiments of the subject technology. Although the methods and processes described herein may be shown with certain steps and operations in a particular order, additional, fewer, or alternative steps and operations performed in similar or alternative orders, or in parallel, are within the scope of various embodiments unless otherwise stated. The method 600 may be implemented by a system such as, for example, client synchronization service 156 of FIG. 2, running on a client device.

11401 The system may receive a set of operations configured to converge a server state and a file system state at operation 620. The set of operations may be, for example, the client set of operations, the server set of operations, or a combined set of operations generated and described with respect to the method 500 of FIG. 5.

11411 At operation 650, the system identifies one or more violations in the set of operations based on a set of rules. The set of rules may be stored by client synchronization service 156 in FIG. 2 and specify a number of constraints, invariants, or conflicts for operations that are to be resolved. The set of rules may be applied to the tree data structures and help control sync behavior. Each rule in the set of rules may also be associated or otherwise linked to a resolution to a violation of that rule. For example, the resolution may include an alteration of one or more operations in the set of operations, a removal off one or more operations, an addition of one or more operations, one or more additional actions to the server state or the file state, or a combination of actions.

11421 For each operation in a set of operations, the system may determine whether any rule in the set of rules is violated. If a rule is violated, the system identifies a resolution of the violation and, at operation 655, performs the resolution. The resolution may include actions such as modifying one or more operations in the set of operations, a removing or adding one or more operations, or additional actions on the server state or the file state.

11431 Once the resolution actions are performed, the system may generate a resolved or rebased set of operation based on the resolution and the set of operations at operation 660 and, at operation 665, provide the resolved set of operations to the appropriate entity for execution. For example, the resolved set of operations may be provided to scheduler 230 of client synchronization service 146 in FIG. 2 for managed execution. Alternatively, if the set of operations is a client set of operations, the resolved set of operations may be provided to the client device. If the set of operations is a server set of operations, the resolved set of operations may be provided to the content management service. Additionally, the method 600 of FIG. 6 may be performed on client set of operations and server set of operations in sequence, in parallel, or in various different orders.

11441 According to some embodiments, each type of operation may be associated with the same or a different set of rules. For example, operation types may include, for example, adding a content item, deleting a content item, editing a content item, moving a content item, renaming a content item, etc. The set of operations may consist of operations each belonging to one of the operation types above. Each operation type maybe associated with a specific set ofrules.

11451 For illustrative purposes, a set of rules for an "Add" operation type may include rules such as file identifiers for content items must be unique in a tree (e.g., no two nodes in a tree may have the same file identifier), a directory file identifier ("DirFilelD") specifying the file identifier of a parent node of the content item must exist in the opposite tree data structure, and a DirFilelD and file name combination for a content item are not used in the opposite tree.

11461 Opposite tree, as used here, refers to the tree data structure that represents the state of the opposing entity. For example, a client set of operations configured to operate on the client device and the resulting changes to the file system on the client device will be reflected in the local tree. Accordingly, the opposite tree for the client set of operations is the remote tree. Similarly, a server set of operations is configured to be transmitted to the content management system to be executed and the resulting changes to the server state will be reflected in the remote tree. Accordingly, the opposite tree for the server set of operations is the local tree.

11471 FIG. 7 shows an example of tree data structures illustrating a violation of a rule for an add operation, in accordance with various embodiments. The tree data structures include remote tree 710, sync tree 750, and local tree 770. When referencing the local tree 770, the remote tree 710 may be considered the opposite tree. On the other hand, when referencing the remote tree 710, the local tree 770 may be considered the opposite tree. FIG. 7 illustrates a set of operations adding the content item represented by node 712 in remote tree 710. For example, a client synchronization service may compare remote tree 710 with sync tree 750, identify the differences, and generate a set of operations that includes the addition of node 712. Node 712 is associated with a FilelD of 4, a DirFilelD of 3 (which references parent node 714, which is node 712's parent), and a file name of "Hi." Parent node 714 is associated with a FilelD of 3, a DirFilelD of1 (which references root node 716, which is node 714's parent), and a file name of "Foo."

11481 The client synchronization service may perform the method 600 of FIG. 6 and determine that the add operation for node 712 violates the "a directory file identifier ("DirFilelD") of the content item must exist in the opposite tree data structure" rule for "add" operation types. This is illustrated in FIG. 7 by the local tree 770 not having a node with a file ID of 3, which references parent node 714 of node 712. This may occur when, for example, after differences between remote tree 710 and sync tree 750 are determined and a set of operations is generated, the "Foo" node corresponding to node 714 is removed from the opposite tree.

11491 The resolution associated for this rule may include deleting the node missing from local tree 770 from sync tree 750 to synchronize sync tree 750 and local tree 770 and rediffing (e.g., finding the difference between) remote tree 710 and sync tree 750. In the scenario illustrated in FIG. 7, node 754 in sync tree 750 would be removed 758 and diffing operations would commence to identify differences between remote tree 710 and sync tree 750. This would result in the inclusion of an add operation of node 714 as well as an add operation for node 712 in the set of operations.

[150] Similarly, a violation of the "file identifiers for content items must be unique in a tree" rule for "add" operation types may be resolved by operations including requesting, from the content management system, a new file ID for the node being added and using the new file ID when adding the node. A violation of the "DirFilelD and file name combination for a content item are not used in the opposite tree" rule for "add" operation types may be resolved by operations including checking via the metadata associated with the two nodes whether the content items are the same. If the content items are the same, it is likely that the content item being added has already been added in other actions. If the content items are not the same, the file name for the content item being added can be renamed. For example, the file name for the content item being added can be appended with the text "(conflicted version)."

Incremental Planner

[151] Although the various tree data structures shown in FIGS. 3, 4, and 7 contain a relatively small number of nodes and are relatively simple in structure, the tree data structures supported by the system may be much larger and complex with multiple levels and potentially large number of nodes at each level. Accordingly the memory usage required to store the tree data structures during operation may be quite large and the computing time and resources required to operate on the tree data structures may be quite large. For example, finding differences between a remote tree and a sync tree and/or a local tree and the sync tree and generating operations needed to converge the remote tree and the sync tree and/or the local tree and the sync tree may require a large amount of memory, time, and other computing resources.

[152] Unfortunately, these computing resources are limited. For example, a client device may have a limited amount of available memory and the length of time needed to diff trees and generate operations may hinder the usability of the client device, the client application, or the content management service provided by the content management system. Furthermore, the more time needed to converge the server state and the file system state, the more likely that intervening changes to either state may render the set of operations being computed or executed and/or the target sync state out of date. Accordingly, various embodiments of the subject technology provide additional technical improvements by incrementally converging the server state and the file system state along with the tree data structures that represent them.

11531 FIG. 8 shows an example method 800 for incrementally converging a server state and a file system state, in accordance with various embodiments of the subject technology. Although the methods and processes described herein may be shown with certain steps and operations in a particular order, additional, fewer, or alternative steps and operations performed in similar or alternative orders, or in parallel, are within the scope of various embodiments unless otherwise stated. The method 800 may be implemented by a system such as, for example, client synchronization service 156 of FIG. 2, running on a client device.

11541 At operation 805, the system may receive modification data that may be used to update either a remote tree or a local tree. For example, server modification data may be received from a content management service and specify modifications or other actions (e.g., an edit, add, delete, move, or rename) associated with one or more content items stored by the content management system. The server modification data may be used to update the remote tree, which represents the server state of content items stored by the content management system. Similarly, client modification data may be received from the client device (e.g., a client application) and specify modifications or other actions associated with one or more content items stored on the client device. The client modification data may be used to update the local tree, which represents the file system state of content items stored on the client device.

11551 Based on the received modification data specifying modifications associated with content items, the system may identify nodes that correspond to the modified content items and add the nodes to a list of modified content items (e.g., add the file identifier associated with the nodes to the list of modified content items) at operation 810. Operations 805 and 810 may continuously occur for some time before the system proceeds to the next stage of the method 800. For example additional modification data may be received and used to update the trees managed by the system and add nodes to the list of modified content items.

11561 In order to incrementally converge the server state and the file system state, the system takes each node in the list of modified content items and determines how the node was modified (e.g., which actions are associated with the node) at operation 815. In some embodiments, the modification data may specify the modification to the node. However, in other embodiments, the system may determine the modifications to the node based on a comparison of the remote tree with the sync tree and/or a comparison of the local tree with the sync tree. For example, the modifications may include the addition of the node, the deletion of the node, the editing of the node, or the moving of the node.

[157] For each node or file identifier for the node in the list of modified content items, the system may perform a series of checks to determine what, if any, modifications were performed on the node. For example, the system may determine whether the file identifier is in the sync tree but not in the remote tree. A file identifier in the sync tree that is not found in the remote tree may indicate that the node has been deleted from the server state that is represented by the remote tree. Accordingly, the client synchronization service may determine that a delete modification on the node has occurred on the remote tree. Similarly, the system may also determine whether the file identifier is in the sync tree but not in the local tree. A file identifier in the sync tree that is not found in the local tree may indicate that the node has been deleted from the file system state that is represented by the local tree. Accordingly, the client synchronization service may determine that a delete modification on the node has occurred on the local tree.

[158] To determine whether an edit modification has been performed on the node the system may compare the metadata for the node in the sync tree with the metadata for the corresponding node (e.g., the node with the same file identifier) in the remote tree and/or the local tree. The metadata may include information that may be used to determine whether the content item represented by the node has been edited. For example, the metadata may include one or more hash values that are generated based on the data in the content item or a portion thereof. The metadata may additionally or alternatively include a size value, a last modified value, or other value for the content item. If the metadata do not match, an edit of the content item may have been edited in the server state represented by the remote tree and/or the file system state represented by the local tree. Accordingly, the system may determine that an edit action has occurred for the node on the remote tree and/or the local tree.

[159] To determine whether the node in the remote tree has been moved, the system may compare the location for the node in the sync tree with the location for the corresponding node (e.g., the node with the same file identifier) in the remote tree and/or the local tree. The location may include, for example, a path where the node is located, a file name, and/or a directory file identifier ("DirFilelD") specifying the file identifier of the node's parent. If the locations match, no move may have occurred. On the other hand, if the locations do not match, a move of the content item may have occurred in the remote tree or the local tree. Accordingly, the client synchronization service may determine that a move action has occurred for the node on the remote tree and/or the local tree.

11601 To determine whether a node has been added to the remote tree, the system may determine if the file identifier in the list of modified content items is in the remote tree or in the local tree, but not in the sync tree. If the file identifier is found in the remote tree or the local tree and not found in the sync tree, the system may determine that an add modification for this node has occurred.

11611 Once the one or more modifications to the nodes in the list of modified content items are determined, the system may determine whether any of those modifications have dependencies at operation 820. As will be illustrated further with respect to FIG. 9, a modification on a node has a dependency when, for example, the modification cannot execute without another modification occurring first.

11621 If the modification does not have a dependency, the system adds the modification to an unblocked list of actions at operation 825. If the modification has a dependency, the modification is blocked for the time being at operation 830 and cannot be executed without another modification being processed first. After each of the modifications are processed, the system may clear the file identifiers associated with the modifications from the list of modified content items.

11631 FIG. 9 shows an example of tree data structures, in accordance with various embodiments. The tree data structures shown in FIG. 9 may be stored at the client device and managed by a system such as client synchronization service 156 in FIG. 2. For the purpose of illustration, only remote tree 910 and sync tree 950 are shown in FIG. 9 and described. Similar operations and description may also be applied to a local tree as well.

11641 Remote tree 910 includes root node 912 with a file identifier of 1, node 914 with a file identifier of 5 and file name of "Foo," node 916 with a file identifier of 6 and file name of "Bar," and node 918 with a file identifier of 7 and file name of "Bye." Sync tree includes root node 952 with a file identifier of 1.

11651 Based on the tree data structures shown in FIG. 9, the system may have identified that nodes with file identifiers of 5, 6, and 7 have been modified at operation 810 and added the nodes to the list of modified content items, as illustrated by reference 980 in FIG. 9. At operation 815, the system determines the list of modifications to nodes in the list of modified content items. As is seen by the comparison of remote tree 910 and sync tree 950, nodes 914, 916, and 918 have been added to remote tree 910. More specifically, as illustrated by reference 982 in FIG. 9, node 916 with file identifier 6 and name "Bar" has been added as a child to node 914 with file identifier 5. This is represented by the "Add(6, 5, Bar)" entry in reference 982. Node 918 with file identifier 7 and name "Bye" has been added as a child to node 914 with file identifier 5. This is represented by the "Add(7, 5, Bye)" entry in reference 982. Node 914 with file identifier 5 and name "Foo" has been added as a child to root node 912 with file identifier 1. This is represented by the "Add(5, /root, Foo)" entry in reference 982.

11661 At operation 820, the system determines that the add modification of node 914 does not have a dependency and, as a result, is unblocked. Accordingly, the system adds the modification associated with node 914 (e.g., the modification represented by the "Add(5, /root, Foo)") entry in reference 982) to an unblocked list of actions at operation 825. This is seen in references 984 in FIG. 9. On the other hand, the modifications for nodes 916 and 918 represented by the "Add(6, 5, Bar)" and the "Add(7, 5, Bye)" entries in reference 982 are dependent on the modification represented by the "Add(5, /root, Foo)" occurring first. In other words, node 916 and/or node 918 cannot be added until node 914 is added. Accordingly, these modifications are included in a blocked list of actions illustrated by reference 986 in FIG. 9.

11671 Returning to the method 800 of FIG. 8, at operation 835, the system may select a set of modifications from the unblocked list of actions and generate a set of operations based on the selected set of modifications. The set of operations is configured to converge the server state and the file system state. The set of operations generated depends on the selected set of modifications from the unblocked list. For example, if the selected set of modifications includes the add modification associated with node 914 (e.g., the modification represented by the "Add(5, /root, Foo)") entry in reference 984) in FIG. 9, the generated set of operations may include retrieving the added content item from the content management system and adding it to the local file system of the client device.

11681 According to some embodiments, the system may select all modifications from the unblocked list of actions to generate one or more sets of operations. However, in some scenarios, the number of modifications in the unblocked list may be quite high and the computing resources (e.g., memory and processing time) needed to process all of the modifications is substantial. In order to reduce these technological burdens, the system may select a smaller set of the modifications in the unblocked list of actions in order to process incrementally. For example, the system may select the first or top X number or percent of modifications to generate operations. In further iterations of the process, the remaining modifications in the unblocked lists maybe processed.

11691 In some embodiments, the modifications in the unblocked list may be ranked for processing. The modifications may be ranked based on, for example, a modification type (e.g., delete modifications are prioritized over add modifications), metadata associated with the modification (e.g., add modifications of content items of smaller size are prioritized over add modifications of content items of larger size, delete modifications of content items of larger size are prioritized over delete modifications of content items of smaller size, etc.).

11701 These rank rules may be stored by the system and may be designed to achieve various performance goals for content synchronization. For example, delete modifications may be prioritized over add modifications in order to free as much of potentially limited storage space for a user before new content items may be added. Adding of smaller content items may be prioritized over larger content items in order to provide as much progress with respect to the number of content items added as soon as possible.

11711 At operation 835, the system may provide the set of operations to the content management system and/or the client device. As noted above, modifications associated with actions performed by the content management system may not be reflected at the client device. Accordingly, in this scenario, the system may generate a client set of operations configured to operate on the content items stored on the client device to converge the server state and the file system state and this client set of operations may be provided to the client device for execution at operation 835.

11721 On the other hand, modifications associated with actions performed by the client device may not be reflected at the content management system. Accordingly, in this scenario, the system may generate a server set of operations configured to operate on the content items stored by the content management system to converge the server state and the file system state and this server set of operations may be provided to the content management system for execution at operation 835.

11731 In some cases, both cases may be true and a client set of operations and a server set of operations may be generated and provided to their intended recipients at operation 835. The set of operations may also include a number of checks to ensure tree constraints are maintained. For example, the set of operations may resolve various conflicts or constraints as discussed with respect to FIG. 6.

11741 Once the set(s) of operations are provided to the intended recipient(s), the method may return to operation 805 and await new modification data. For example, with respect to the scenario illustrated in FIG. 9, the set of operations may include retrieving the content item associated with node 914 from the content management system and adding it to the local file system of the client device. This would result in the addition of a node corresponding to node 914 in the local tree (not shown in FIG. 9) and sync tree 950. On the next iteration of process 800 of FIG. 8, the add modifications of node 916 and node 918 represented by the "Add(6, 5, Bar)" and the "Add(7, 5, Bye)" entries in reference 982 are no longer blocked because their parent, node 914, has already been added to the sync tree. Accordingly, the add modifications of node 916 and node 918 represented by the "Add(6, 5, Bar)" and the "Add(7, 5, Bye)" entries in reference 982 may be added to the unblocked list of actions and used to generate one or more sets of operations configured to converge the server state and the file system state.

11751 The set(s) of operations may provide one or more steps for the incremental convergence of the server state and the file system state. Although implementing an incremental process may be more complex at times, the incremental process may achieve a reduction in processing time and reduction in the memory required. These and other initial technological improvements naturally lead to additional technological improvements. For example, because processing time is reduced, the likelihood of additional changes from the client device or the content management system making certain modifications obsolete or out of data is reduced as well.

11761 With respect to FIG. 9, various groupings of content items, modifications, actions, or file identifiers are described as lists for the purpose of illustration. Other types of data structures are also compatible. For example, the unblocked list of actions may be implemented as a B-tree data structure in order to keep data sorted and allow searches, sequential access, insertions, and deletions in logarithmic time.

Scheduler

11771 In some embodiments, a client synchronization service may generate a set or sequence of operations configured to converge the server state and the file system state and provide the operations to the content management system or client device for execution. However, in some scenarios, changes on the file system of the client device or on the content management system may cause the generated set of operations to become out of date or obsolete while the set of operations is in the process of executing. Various embodiments are directed to providing a technical solution to these and other technical problems. For example, the client synchronization service may be configured to monitor changes on the file system of the client device or on the content management system and update the client device and/or content management as needed. Furthermore, the client synchronization service may be configured to improve performance and reduce processing times by allowing for concurrent execution of operations.

11781 According to some embodiments, planner 225 of client synchronization service 156 shown in FIG. 2 may generate a plan or plan of operations that consists of an unordered set of operations. All operations within a plan have no dependencies and, as a result, are able to be executed concurrently in separate threads or in any order. The operations in the plan, according to some embodiments, are abstract instructions that may be taken by the content management system and/or the client device in order to converge the states and tree data structures. Example instructions may include a remote or local add of a content item, a remote or local delete of a content item, a remote or local edit of a content item, or a remote or local move of a content item.

11791 Scheduler 230 of client synchronization service 156 shown in FIG. 2 may be configured to receive the plan of operations from planner 225, manage the execution of the operations in the plan, determine if the plan has been updated or changed, and manage the execution of the updated or changed plan. For example, scheduler 230 may coordinate with file system interface 205 and server interface 210 to execute the tasks and steps needed to implement operations in the plan. This may include receiving confirmations from the file system or content management system or error handling activities such as handling retries when there is no network connectivity or when a content item is locked by some other application.

11801 Each operation may be implemented by a script or thread referred to as a task. The task coordinates the application of an associated operation and may include one or more steps needed to implement the operation. For example, a "local add operation" may indicate that a content item has been added to the local file system of the client device and, as a result, the content item should be added at the content management system in order to sync the server state and the file system state. Accordingly, the local add operation may be associated with a "local add task" that includes one or more steps needed to implement the local add operation. The steps may include one or more of notifying the content management system of the new content item, uploading the content item to the content management system in one or more blocks of data, confirming that all blocks of data have been received by the content management system, making sure the content item is not corrupted, uploading metadata for the content item to the content management system, and committing the adding of the content item to the appropriate location at the content management system.

11811 A task may begin execution, suspend at well-defined points while waiting on the completion of other events, resume when the events have occurred, and eventually terminates- According to some embodiments, scheduler 230 is configured to cancel, regenerate, or replace tasks. For example, based on changes to the server state or the file system state, a task may become stale before it is executed and scheduler 230 may cancel the stale task before it is executed.

11821 As described above, planner 225 may generate a plan of operations based on a set of tree data structures (e.g., a remote tree, a sync tree, and a local tree). Over time, planner 225 continues to generate plans of operations based on the status of the tree data structures. If the tree data structures change to reflect the state of the server state and the file system state, planner 225 may also generates a new updated plan that differs from a previous plan. Scheduler 230 executes each plan of operations generated by the planner 225.

11831 In some scenarios, changes in the operations of a subsequent plan may cause unintended behaviors conflict with an operation in the previous plan that is in the process of execution. For example, as operations in a first plan are being executed, one or more of the operations are canceled (or are not present) in the second plan. To illustrate, FIG. 10 shows an example scenario in which, at time tl, the server state represented by the remote tree and the file system state represented by the local tree are synchronized as shown by the remote tree, the sync tree, and the local tree all matching. Based on this synchronized state, planner 225 may generate a plan with no operations (e.g., an empty plan) at tl.

11841 A user on the client device may delete content item A from the local file system or move content item A out of a folder managed by client synchronization service 156, which is reflected by the removal of node A from the local tree at time t2. Planner 225 may generate a plan that includes operation LocalDelete(A) based on the state of the tree data structures at time t2. Scheduler 230 may initiate the task or steps required to implement the LocalDelete(A) operation. These steps may include transmitting instructions to the content management system to delete content item A.

11851 After instructions to delete content item A are transmitted to the content management system, the user on the client device may undo the delete of content item A or move content item A back to the previous location. The local tree is updated based on this new action at time t3 and planner may generate a new plan that is empty with no operations. Once again, the tree data structures match and the system is in a synchronized state at time t3.

11861 However, because instructions to delete content item A were transmitted to the content management system, the content management system deletes content item A from the server state. Although scheduler 230 may attempt to cancel the deletion of content item A, the instructions may have already been transmitted and completed by the content management system. This change in the server is communicated to client synchronization server 156, which updates the remote tree by deleting node A at time t4. Planner 225 could notice the change in the remote tree and the difference between the remote tree and the sync tree and determine that content item A was removed at the server state. Accordingly, planner 225 would create a plan with a RemoteDelete(A) operation at time t4. In an effort to synchronize the server state and the file system state, content item A would eventually be deleted from the client device and the local tree.

11871 Problematically, the removal of content item A from the server state, the generation of the RemoteDelete(A) operation, and the eventual removal of content item A from the file system state are all not intended and may cause further problems down the line for the user. Furthermore, in some cases, applications or processes may also access content items and unintentional synchronization behavior may cause a cascade of additional technical issues. Various embodiments are directed to preventing unintended consequences in synchronization of content items between a server state and a file system state.

11881 According to some embodiments, when canceling a task for a stale operation that is no longer in a plan of operations, scheduler 230 may wait for the cancelation to be completed before proceeding to initiate the execution of other tasks. For example, scheduler 230 may wait to receive confirmation of the cancelation from the client device or the content management system before proceeding with other tasks. Scheduler 230 may determine whether the task has been initiated and if the task has not been initiated, scheduler may cancel the task and confirm that the task is no longer awaiting execution. If the task has been initiated, the confirmation may come from the client device or the content management system and notify the scheduler that all of the steps associated with the canceled task have been undone. According to some implementations, scheduler 230 does not allow for cancelation of a task once it has been initiated. This may be the case for all tasks or a certain subset of tasks or task types (e.g., a commit task that sends an update on the file system state to the content management system for synchronization with the server state).

11891 In order to improve performance and allow for concurrent execution of tasks as well as the cancelation of tasks, scheduler 230 may also be configured to manage the execution and cancelation of tasks based on differences between a first plan of operations and an updated second plan of operations. FIG. 11 shows an example Venn diagram 1100 representation of two plans of operations, in accordance with various embodiments of the subject technology. Planner 225 may generate a plan 1 1110 with a first set of operations, receive an update to the tree data structures, and generate an updated plan 2 1120 with a second set of operations.

11901 Plan 1 1110 and plan 2 1120 may share a number of common operations, which is represented by portion 1130 of the Venn diagram 1100. Plan 1 1110 and plan 2 1120 may also share a number of operations that are not in common. For example, operations in plan 1 1110 that are not in plan 2 1120 are stale and no longer current based on the update to the tree structures detected by planner 225. These stale operations of plan 1 1110 are represented by portion 1140 of Venn diagram 1100. New operations in plan 2 1120 that are not in plan 1 1110 are represented by portion 1150. Each of portions 1130, 1140, and 1150 which represent the differences and commonalities between plan 1 1110 and plan 2 1120 may include no operations or many operations depending on the updates to the server state and the file system state that are reflected in the tree data structures.

11911 Because the operations in portion 1140 are no longer in the most recent plan, scheduler 230 may cancel tasks associated with these operations. In order to prevent unintended synchronization behavior, tasks associated with operations in plan 2 that are not in plan 1 (e.g., in portion 1150) are postponed until the cancelation of tasks associated with operation in portion 1140 is completed. However, because operations in each plan are configured to be able to be executed concurrently, tasks associated with operations in the intersection of plan 1 and plan 2 represented by portion 1130 may be executed concurrently with the cancelation of tasks associated with operation in portion 1140 without having to wait for their completion. By allowing for the concurrent cancelation of task associated with portion 1140 and the execution of tasks associated with portion 1130, more efficient use of available computing resources may be achieved as well as a reduction in processing time.

11921 FIG. 12 shows an example method for managing changes in plans of operations, in accordance with various embodiments of the subject technology. Although the methods and processes described herein may be shown with certain steps and operations in a particular order, additional, fewer, or alternative steps and operations performed in similar or alternative orders, or in parallel, are within the scope of various embodiments unless otherwise stated. The method 1200 may be implemented by a system such as, for example, client synchronization service 156 of FIG. 2, running on a client device.

11931 The system may be configured to receive updates from a content management system and/or the client device with regards to content items associated with a content management service. For example the system may receive server modification data for content items stored by a content management service and update, based on the server modification data, a remote tree. The remote tree represents the server state for content items stored by the content management system. The system may also receive client modification data for content items stored on the client device and update, based on the client modification data, a local tree. The local tree represents the file system state for content items stored on the client device.

11941 At operation 1205, the system may receive a first set of operations configured to converge a server state associated with the content management system and a file system state associated with the client device. For example, the system may identify differences between a sync tree and a remote tree or the sync tree and a local tree and generate the first set of operations based on any differences between the trees. The sync tree represents a known synced state between the server state and the file system state.

11951 The system may begin to implement the first set of operations. For example, in some cases, the operations are in a format ready to be transmitted to the content management system and/or the client device for execution. In other cases, the operations may be translated into one or more tasks, scripts, or execution threads that may be managed by the system. The system may interface with the content management system and/or the client device according to the tasks, scripts, or execution threads in order to converge the server state and the file system state.

11961 During this time, the system may continue to receive modification data from a content management system and/or the client device with regards to content items associated with the content management service. Based on the modification data, the system may update the remote tree or local tree and generate a second set of operations based on the updates to the tree data structures. At operation 1210, the system may receive the second set of operations.

11971 At operation 1215, the system identifies a first operation in the first set of operations that is not in the second set of operations, if any. If the system finds an operation in the first set of operations that is not in the second set of operations, this operation may be stale and out of date as a result of changes specified in the modification data. Accordingly, the system will initiate the cancelation of the first operation at operation 1220. The cancelation of the first operation may include a number of steps, a number of confirmation receipts for the steps, and a non-trivial amount of processing time.

11981 At operation 1225, the system identifies a second operation that is included in both the first set of operations and the second set of operations, if any. If the system finds an operation in both the first set of operations and the second set of operations, this operation may be still be valid notwithstanding changes specified in the modification data. Furthermore, since the operations in both sets of operations are configured to be able to be executed concurrently or in any order with respect to other operations in the set, the second operation can continue execution while the first operation is canceled. Accordingly, the system will initiate the execution of the second operation at operation 1230 without waiting for the first operation to complete cancelation.

11991 At operation 1235, the system identifies a third operation that is in the second set of operations, but not in the first set of operations, if any. If the system finds an operation in the second set of operations that is not in the first set of operations, this operation may be a new operation as a result of changes specified in the modification data. In order to prevent unintended consequences, the system will initiate the wait for the completion of the cancelation of the first operation. At operation 1240, the system may determine that the first operation has completed cancelation and, as a result, initiate the execution of the third operation at operation 1245.

Updating the Local Tree

[200] As described above, the local tree is configured to reflect the file system state for content items stored on the local file system of the client device. For example, file system interface 205 of client synchronization service 156 in FIG. 2 is configured to make changes to the local file system of the client device (e.g., add, delete, move, edit, or rename one or more content items), detect changes to the local file system, and update the local tree based on the changes to the local file system. The changes may be caused by a user action on the file system, a third-party application running on the client device, or by the client synchronization service synchronizing the file system state with the server state.

[201] Various embodiments of the subject technology provide various technical solutions to updating the local tree based on changes to the local file system. The local tree, along with the other tree data structures, is crucial to the synchronization processes between the client device and the content management system in various embodiments. For example, once an update to the local tree is made, the rest the system reacts to the update and, in some cases, the changes to the local tree may be synchronized and applied to the server state at the content management system. Accordingly, it is important to be careful about how the local tree is updated.

[202] For example, if a user renames a file from A.txt to B.txt, in some cases, the system may detect a delete of content item A.txt and an add of content item B.txt. This may cause a node for A.txt to be deleted on the local tree and a node for B.txt to be added. However, this results in a case where, for some time, no node for the renamed content item exists on the local tree. This can cause significant damage to data integrity because the client device, the client application, and/or the client synchronization service may be shut down, fail, or reboot before the node for B.txt is added and, as a result, a user's content item is lost. The loss of the user's content item may then be synchronized to the server state at the content management system. Similar risks are associated with a user moving a content item from one location to another.

[203] Additionally, the changes to the local file system may be detected out of order and may include a large number of changes that are not necessarily all related to a single action by a user or application. The client application may also be turned off or not running while many changes to the local file system are made. On startup, the client application may crawl the local file system, compare it with the local tree, and determine which changes to the local file system have occurred while the client application was off. These changes may not be in proper chronological order. These factors may also result in unintended synchronization behavior if the local tree is not carefully updated.

12041 A set of constraints may be used to ensure tree data structure integrity and protect against unintended synchronization behavior. The constraints may include, for example, that (1) all nodes in a tree are associated with afile identifier (filelD), (2)

all nodes in a tree have a unique file identifier, (3) non-empty parent nodes cannot be deleted, (4) deleted nodes are actually deleted (and not merely moved) or removed from a location managed by the client synchronization service, (5) all sibling nodes have unique file names irrespective of case, (6) all nodes must have an existing parent, and/or (7) all sibling nodes agree on their parents file id (DirFilelD). In some implementations, a subset of the constraints above may be used, alternative or additional constraints may be used, or a combination. The set of constraints may be applied to all tree data structures or merely a subset of the tree data structures while a different set or sets of constraints may be applied to other tree data structures.

12051 When a change to the local file system is detected, the change may be checked against the set of constraints. If the change is consistent with the set of constraints, the local tree can be updated based on the change to the local file system. If the change violates one of the constraints, the constraint may require additional conditions to be satisfied. For example, a constraint may require additional file events to be observed before the changes can be applied to the local tree, one or more remediation steps to be performed, or a combination. As actions occur to satisfy certain constraints (e.g., remediation steps are taken or additional file events occur) other constraints may be violated. Accordingly, the set of constraints may be continually checked until all constraints are satisfied. Once the constraints are satisfied, the changes associated with the file events may be applied to the local tree.

12061 FIG. 13 shows an example scenario, in accordance with various embodiments of the subject technology. In particular, FIG. 13 shows the current state of local tree 1310 when file event 1315 is detected. For example, the client synchronization service may compare the file system with local tree 1310 and discover that a content item exists in the file system at the path /root/a/b/c.txt but a node for a content item at /root/a/b/c.txt does not exist in local tree 1310. Accordingly, a file event 1315 may be generated specifying an add of a node at /root/a/b/c.txt is needed on local tree 1310.

12071 The client synchronization service may add file event 1315 to a set of observed file events for update and determine whether the observed file event 1320 is consistent with a set of constraints and discover that observed file event 1320 violates one of the constraints in the set. In the scenario illustrated in FIG. 13, the observed file event 1320 violates the "all nodes must have an existing parent" constraint. More specifically, the parent of the node to be added at /root/a/b/c.txt does not exist nor does the grandparent of the node. Accordingly, additional file events (the addition of the parent and grandparent node) must be observed before the change is applied to the local tree.

12081 The client synchronization service may detect additional file events and add them to the set of observed file events for update. For example, the client synchronization service may detect the add /root/a file event and the add /root/a/b file event. Once these file events are observed, the violated constraint is satisfied (and no other constraints are violated). As a result, all of the observed file events for update may be applied to the local tree. More specifically, a node may be added at /root/a, a node may be added at /root/a/b, and a node may be added at /root/a/b/c.txt. Accordingly, the client synchronization service groups together related file events for an atomic or unitary update. As will be described in further detail, grouping together related file events for an atomic update to the local tree increases tree data structure integrity, protects against unintended synchronization behavior, and prevents intermediate states in the local tree.

12091 FIG. 14 shows an example method for updating a local tree, in accordance with various embodiments of the subject technology. Although the methods and processes described herein may be shown with certain steps and operations in a particular order, additional, fewer, or alternative steps and operations performed in similar or alternative orders, or in parallel, are within the scope of various embodiments unless otherwise stated. The method 1400 may be implemented by a system such as, for example, client synchronization service 156 of FIG. 2 running on a client device.

[210] At operation 1405, the system detects a file event and adds the file event to a set of observed file events. For example, on startup, the system may crawl the file system of the client device, collect information on the file system, and compare the collected information with the local tree which represents the last known state of the file system. The system may identify differences between the local tree and the file system of the client device and generate a number of file events based on the identified differences. Alternatively, or additionally, the system may monitor the file system during runtime, detect changes made to the file system that are not reflected in the local tree, and generate file events based on the detected changes to the file system. The generated file events may be thought of as observations about the file system that are made by the system.

[211] At operation 1410, the system will check the set of observed file events against a set of local tree constraints to determine whether any of the observed file events violates a local tree constraint. If none of the observed file events violate any of the constraints in the set of local tree constraints are violated, the set of observed file events may be used to update the local tree at operation 1435.

[212] Each violation of a constraint may be associated with a remediation configured to satisfy the constraint. For example, a violation of the "all nodes must have an existing parent" constraint may be associated with a requirement that the addition of the parent node be observed, as illustrated in FIG. 13.

12131 In other cases, a violation of a constraint may require actions to be taken to resolve the violation and satisfy the constraint. For example, when a user copies a content item in the file system and creates a new copy of an existing content item, the new content item may have the same file identifier as the original content item. The system may observe the addition of the new copy, but this new copy violates the constraint that "all nodes in a tree have a unique file identifier." The violation of this constraint may be associated with the remediation steps of requesting a new file identifier for the content item and assigning the new content identifier to the content item, thus resolving the violation and satisfying the constraint, before the local tree is updated. Accordingly, the local tree will at no point be in a state that violates any constraint.

12141 In another example, a user may create a new content item in a location of the file system where the file name already exists, albeit with differing letter case. To illustrate, a user may create a file named "A.txt" when the file "a.txt" already exists with the same file system path. The operating system of the client device may allow for this while the client synchronization service may not. The system may observe the addition of the new content item, but this new content item violates the constraint that "all sibling nodes have unique file names irrespective of case." The violation of this constraint may be associated with the remediation steps of editing the name of the new content item to denote that a case conflict exists. For example, the "A.txt" file may be renamed "A(case conflict).txt," thus resolving the violation and satisfying the constraint. The file event may be removed and the process restarted such that a new file event for the addition of the "A(case conflict).txt" content item is detected or the file event for "A.txt" may be updated to reflect the new name "A(case conflict).txt."

12151 If one or more observed file events violate one or more constraints, the system may determine whether the violated constraint requires remediation actions to be taken at operation 1415 or whether the violated constraint requires additional file events to be observed at operation 1425. If additional remediation actions are required, at operation 1420, the system may execute the additional remediation actions. If additional file events are to be observed, at operation 1430, the system may detect additional file events and add the file events to the set of observed file events at operation 1430.

12161 The process then returns to operation 1410 to determine whether the set of observed file events violate the local tree constraints. In some cases, the execution of the remediation actions or the new file events added to the set of file events may cause new violations of one or more constraints that must be resolved before an update to the local tree can be performed. Accordingly, the process may iterate until no more violations of the local tree constraints exist. The process may then proceed to operation 1435, where the system may update the local tree based on the observed set of file events.

Updating the Local Tree with Move or Rename Changes

12171 According to some implementations, move or rename operations on content items on the local file system may introduce additional technical problems. For example, in some cases when content items such as files or folders are moved from an old location to a new location by the user or an application, the operation may appear to the file system or client application as a delete of the content items from the old location and an add of new content items at the new location. Similarly, a rename of a content item from an old filename to a new file name may appear as a delete of the content item with the old filename and an add of a new content item with the new file name. Furthermore, if the content item is a folder that is the parent of many other content items, with a potentially deep and complex tree structure, a move or rename of the content item may also appear as a delete of all descendent content items from their old location or path to a new location or path.

12181 As described above, intermediate states where content items are deleted or removed from the local tree before they are re-added in the new location or with the new name is undesirable and increases data vulnerability where a user's data may be lost. Additionally, move or rename operations appear as delete operations to the client synchronization service until a corresponding add operation is detected. However, the add operation may not be detected for a long time after the delete operation is detected based on the size and complexity of the local file system. For example, the client synchronization service may crawl one portion of the local file system and discover the delete of the content item and not discover that the content item has been added to another portion of the local file system, thereby completing the move operation, until the client synchronization service crawls that portion of the local file system.

12191 Various embodiments of the subject technology are directed to providing technical solutions to these and other technical problems by providing a more efficient and faster method of determining whether a delete operation is part of a move or rename operation or simply a delete operation.

[220] FIG. 15 shows an example method for updating a local tree in response to a move or rename operation, in accordance with various embodiments of the subject technology. Although the methods and processes described herein may be shown with certain steps and operations in a particular order, additional, fewer, or alternative steps and operations performed in similar or alternative orders, or in parallel, are within the scope of various embodiments unless otherwise stated. The method 1500 may be implemented by a system such as, for example, client synchronization service 156 of FIG. 2 running on a client device.

[221] At operation 1505, the system detects a delete event for a content item. The system may crawl or monitor changes to the local file system of the client device or a portion of the local file system that the system is configured to manage (e.g., a content management folder). The system may compare the local file system with the local tree in order to identify differences between the local file system and the local tree. The delete event may be detected based on one or more identified differences. For example, a node for a content item exists in the local tree at a particular location but does not exist at that location on the local file system. This may indicate that a user or application has performed an action that caused the content item to be removed from that location, which causes the system to detect the delete event.

[222] The action that caused the delete event to be detected may be caused by a user or application moving the content item to another location monitored by the system, moving the content item to another location not monitored by the system, renaming the content item (which may be treated as a move by some file systems), or actually deleting the content item. In order to determine what user action caused the delete event and/or whether or not an add event associated with the delete event will be or has already been detected, the system may identify an operating system provided identifier for the content item at operation 1510.

[223] In some embodiments, the operating system provided identifier may be an inode identifier and is different from the file identifier provided by the content management system and/or client synchronization service. In many cases, the operating system may provide the inode identifier in order to, among other things, allow for quick querying of the location of a content item based on the inode identifier. For example, some operating systems may provide an interface where the system may query for a current path or location of a content item using the inode identifier as a key. At operation 1515, the system may determine the location of the content item by querying the operating system for the location of the content item. In response to the query, the operating system may return with the current location or path in the local file system of the content item referenced by the inode identifier.

[224] Using the current location of the content item, the system may determine what action caused the delete event. For example, if the current location is a null location or otherwise indicates that the content item is no longer on the local file system, the action that caused the delete event is an actual delete. Accordingly, the system can appropriately delete the node for the content item from the local tree. If the current location is a location not managed by the system (e.g., the client synchronization service), the action that caused the delete event is likely a move of the content item from its previous location to its current location. However, because the content item is moved outside the territory managed by the system, the system no longer needs to track the content item and can delete the node for the content item from the local tree.

12251 If the current location is a new location that is still managed by the system, the action that caused the delete event is also a move of the content item from its previous location to its current location. However, because the content item is still within a territory managed by the system, the system should await the detection of a corresponding add event and treat the delete event and the add event together as a move action and update the local tree atomically, mirroring the actual action that caused the delete event.

12261 Similarly, if the current location is the same location as the old location, which is managed by the system, the action that caused the delete event is also a rename of the content item from its previous location to its current location. In some file systems, rename operations and move operations are related in that a rename operation is treated as a move operation from one location with one name to the same location with a new name. Accordingly, the system should await the detection of a corresponding add event (with the new name) and treat the delete event and the add event together as a move or rename action and update the local tree atomically, mirroring the actual action that caused the delete event.

12271 Accordingly, at operation 1520, the system determines whether the delete event is associated with an add event for the content item based on the location of the content item. If the delete event is not associated with an add event, the delete event may be processed at operation 1525. If the delete event is associated with an add event, the system may wait for the add event, detect the add event for the content item at operation 1530, and process the delete event with the add event in a unitary update to the local tree at operation 1535.

12281 Although method 1400 of FIG. 14 and method 1500 of FIG. 15 are described separately, the two methods may work in conjunction with one another in order to update the local tree. For example, if the delete event is not associated with an add event, the delete event may be processed without combining the delete event with a corresponding add event at operation 1525 of FIG 15. According to some embodiments, processing the delete event may include operations illustrated in FIG. 14 where, for example, the delete event may be added to a set of observed file events and checked to determine whether a local tree constraint is violated.

[229] For example, if the content item associated with the delete event has one or more descendant nodes in the local tree, the "non-empty parent nodes cannot be deleted" constraint may be violated. The remediation for this violation may include waiting to observe additional file events (e.g., delete events for every descendant node of the content item). Once the additional file events are detected and constraints for these additional file events may be checked, including a check to determine if additional delete file events are associated with additional corresponding add events. Once all of the observed file events are validated, the file events may be batched together and used to update the local tree.

[230] Similarly, if the delete event is associated with an add event, the system may wait for the add event. Processing the delete event with the add event in a unitary update to the local tree at operation 1535 may include adding both events to the set of observed file events, determining whether they violate any local tree constraints, performing appropriate remediations if they do, and updating the local tree based on the entire set of observed file events.

[231] FIG. 16 shows an example of computing system 1600, which can be for example any computing device making up client device 150, content management system 110 or any component thereof in which the components of the system are in communication with each other using connection 1605. Connection 1605 can be a physical connection via a bus, or a direct connection into processor 1610, such as in a chipset architecture. Connection 1605 can also be a virtual connection, networked connection, or logical connection.

[232] In some embodiments computing system 1600 is a distributed system in which the functions described in this disclosure can be distributed within a datacenter, multiple datacenters, a peer network, etc. In some embodiments, one or more of the described system components represents many such components each performing some or all of the function for which the component is described. In some embodiments, the components can be physical or virtual devices.

[233] Example system 1600 includes at least one processing unit (CPU or processor) 1610 and connection 1605 that couples various system components including system memory 1615, such as read only memory (ROM) 1620 and random access memory (RAM) 1625 to processor 1610. Computing system 1600 can include a cache ofhigh speed memory 1612 connected directly with, in close proximity to, or integrated as part of processor 1610.

12341 Processor 1610 can include any general purpose processor and a hardware service or software service, such as services 1632, 1634, and 1636 stored in storage device 1630, configured to control processor 1610 as well as a special-purpose processor where software instructions are incorporated into the actual processor design. Processor 1610 may essentially be a completely self-contained computing system, containing multiple cores or processors, a bus, memory controller, cache, etc. A multi-core processor may be symmetric or asymmetric.

12351 To enable user interaction, computing system 1600 includes an input device 1645, which can represent any number of input mechanisms, such as a microphone for speech, a touch-sensitive screen for gesture or graphical input, keyboard, mouse, motion input, speech, etc. Computing system 1600 can also include output device 1635, which can be one or more of a number of output mechanisms known to those of skill in the art. In some instances, multimodal systems can enable a user to provide multiple types of input/output to communicate with computing system 1600. Computing system 1600 can include communications interface 1640, which can generally govern and manage the user input and system output. There is no restriction on operating on any particular hardware arrangement and therefore the basic features here may easily be substituted for improved hardware or firmware arrangements as they are developed.

12361 Storage device 1630 can be a non-volatile memory device and can be a hard disk or other types of computer readable media which can store data that are accessible by a computer, such as magnetic cassettes, flash memory cards, solid state memory devices, digital versatile disks, cartridges, random access memories (RAMs), read only memory (ROM), and/or some combination of these devices.

12371 The storage device 1630 can include software services, servers, services, etc., that when the code that defines such software is executed by the processor 1610, it causes the system to perform a function. In some embodiments, a hardware service that performs a particular function can include the software component stored in a computer-readable medium in connection with the necessary hardware components, such as processor 1610, connection 1605, output device 1635, etc., to carry out the function.

12381 For clarity of explanation, in some instances the present technology may be presented as including individual functional blocks including functional blocks comprising devices, device components, steps or routines in a method embodied in software, or combinations of hardware and software.

12391 Any of the steps, operations, functions, or processes described herein may be performed or implemented by a combination of hardware and software services or services, alone or in combination with other devices. In some embodiments, a service can be software that resides in memory of a client device and/or one or more servers of a content management system and perform one or more functions when a processor executes the software associated with the service. In some embodiments, a service is a program, or a collection of programs that carry out a specific function. In some embodiments, a service can be considered a server. The memory can be a non transitory computer-readable medium.

12401 In some embodiments the computer-readable storage devices, mediums, and memories can include a cable or wireless signal containing a bit stream and the like. However, when mentioned, non-transitory computer-readable storage media expressly exclude media such as energy, carrier signals, electromagnetic waves, and signals per se.

12411 Methods according to the above-described examples can be implemented using computer-executable instructions that are stored or otherwise available from computer readable media. Such instructions can comprise, for example, instructions and data which cause or otherwise configure a general purpose computer, special purpose computer, or special purpose processing device to perform a certain function or group of functions. Portions of computer resources used can be accessible over a network. The computer executable instructions may be, for example, binaries, intermediate format instructions such as assembly language, Firmware, or source code. Examples of computer-readable media that may be used to store instructions, information used, and/or information created during methods according to described examples include magnetic or optical disks, solid state memory devices, flash memory, USB devices provided with non-volatile memory, networked storage devices, and so on.

[242] Devices implementing methods according to these disclosures can comprise hardware, firmware and/or software, and can take any of a variety of form factors. Typical examples of such form factors include servers, laptops, smart phones, small form factor personal computers, personal digital assistants, and so on. Functionality described herein also can be embodied in peripherals or add-in cards. Such functionality can also be implemented on a circuit board among different chips or different processes executing in a single device, by way of further example.

[243] The instructions, media for conveying such instructions, computing resources for executing them, and other structures for supporting such computing resources are means for providing the functions described in these disclosures.

[244] Although a variety of examples and other information was used to explain aspects within the scope of the appended claims, no limitation of the claims should be implied based on particular features or arrangements in such examples, as one of ordinary skill would be able to use these examples to derive a wide variety of implementations. Further and although some subject matter may have been described in language specific to examples of structural features and/or method steps, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to these described features or acts. For example, such functionality can be distributed differently or performed in components other than those identified herein. Rather, the described features and steps are disclosed as examples of components of systems and methods within the scope of the appended claims.

Claims

CLAIMS What is claimed is:

1. A computer-implemented method comprising: detecting a delete event for a first content item at a first location on a local file system of a client device, wherein the first location on the local file system is managed by a content management system; determining that the delete event is associated with an add event for the first content item based on querying for a second location for the first content item; and based on the determining: waiting for the add event to be processed; and after the add event is processed, processing the delete event for the first content item at the first location with the add event for the first content item at the second location in a unitary update to a local tree, wherein the local tree represents a file system state.

2. The computer-implemented method of claim 1, further comprising: comparing, by the content management system, nodes of the local file system with corresponding nodes of the local tree, resulting in an identification of the delete event.

3. The computer-implemented method of claim 1, further comprising: detecting a second delete event for a second content item at a third location on the local file system of the client device; determining that a current location of the second content item is not found; and deleting a node for the second content item from the local tree.

4. The computer-implemented method of claim 1, wherein the second location is identical to the first location and the first content item associated with the add event has a new name.

5. The computer-implemented method of claim 1, further comprising: identifying a metadata identifier associated with the first content item for the delete event and a same metadata identifier associated with the first content item for the add event, whereby causing the determining that the delete event is associated with the add event.

6. The computer-implemented method of claim 1, wherein the processing the delete event with the add event includes adding the delete event and the add event to a set of observed file events and updating the local tree based on an entire set of observed file events.

7. The computer-implemented method of claim 1, further comprising: determining that an update action associated with the first content item violates a constraint that non-empty parent nodes cannot be deleted; and based on the determining: waiting for delete events for every descendant node of a parent node associated with the first content item to be processed; and after the delete events for the every descendant node of the first content item are processed, processing the delete event for the first content item at the first location with the delete events for the every descendant node of the first content item in the unitary update to the local tree.

8. A non-transitory computer readable medium comprising instructions, the instructions, when executed by a computing system, cause the computing system to: detect a delete event for a first content item at a first location on a local file system of a client device, wherein the first location on the local file system is managed by a content management system; determine that the delete event is associated with an add event for the first content item based on querying for a second location for the first content item; and based on the determining: wait for the add event to be processed; and after the add event is processed, process the delete event for the first content item at the first location with the add event for the first content item at the second location in a unitary update to a local tree, wherein the local tree represents a file system state.

9. The non-transitory computer readable medium of claim 8, wherein the instructions further cause the computing system to: compare, by the content management system, the local file system with the local tree, resulting in an identification of the delete event.

10. The non-transitory computer readable medium of claim 8, wherein the instructions further cause the computing system to: detect a second delete event for a second content item at a third location on the local file system of the client device; determine that a current location of the second content item is not found; and delete a node for the second content item from the local tree.

11. The non-transitory computer readable medium of claim 8, wherein the second location is identical to the first location and the first content item associated with the add event has a new name.

12. The non-transitory computer readable medium of claim 8, wherein the instructions further cause the computing system to: identify a metadata identifier associated with the first content item for the delete event and a same metadata identifier associated with the first content item for the add event, whereby causing the determining that the delete event is associated with the add event.

13. The non-transitory computer readable medium of claim 8, wherein the processing the delete event with the add event include adding the delete event and the add event to a set of observed file events and updating the local tree based on an entire set of observed file events.

14. The non-transitory computer readable medium of claim 8, wherein the instructions further cause the computing system to: determine that an update action associated with the first content item violates a constraint that non-empty parent nodes cannot be deleted; and based on the determining: wait for delete events for every descendant node of a parent node associated with the first content item to be processed; and after the delete events for the every descendant node of the first content item are processed, process the delete event for the first content item at the first location with the delete events for the every descendant node of the first content item in the unitary update to the local tree.

15. A system comprising: one or more processors; and a non-transitory computer-readable medium storing instructions that, when executed by the one or more processors, cause the one or more processors to: detect a delete event for a first content item at a first location on a local file system of a client device, wherein the first location on the local file system is managed by a content management system; determine that the delete event is associated with an add event for the first content item based on querying for a second location for the first content item; and based on the determining: wait for the add event to be processed; and after the add event is processed, process the delete event for the first content item at the first location with the add event for the first content item at the second location in a unitary update to a local tree, wherein the local tree represents a file system state.

16. The system of claim 15, wherein the instructions further cause the one or more processors to: compare, by the content management system, nodes of the local file system with corresponding nodes of the local tree, resulting in an identification of the delete event.

17. The system of claim 15, wherein the instructions further cause the one or more processors to: detect a second delete event for a second content item at a third location on the local file system of the client device; determine that a current location of the second content item is not found; and delete a node for the second content item from the local tree.

18. The system of claim 15, wherein the second location is identical to the first location and the first content item associated with the add event has a new name.

19. The system of claim 15, wherein the instructions further cause the one or more processors to: identify a metadata identifier associated with the first content item for the delete event and a same metadata identifier associated with the first content item for the add event, whereby causing the determining that the delete event is associated with the add event.

20. The system of claim 15, wherein the processing the delete event with the add event include adding the delete event and the add event to a set of observed file events and updating the local tree based on an entire set of observed file events.